Comparison of Different Machine Learning Models in Prediction of Postirradiation Recurrence in Prostate Carcinoma Patients
Table 4
The value of nadir PSA, time to nadir PSA, PSA doubling time, PSA velocity, and PSA in the moment of disease reevaluation, in prediction of disease progression.
Characteristic
The presence/absence of disease progression – training group
Presence
Absence
Wilcoxon rank sum test
Nadir PSA (ROC cut-off value)
≤0.3095
20 (37.0%)
17 (68.0%)
>0.3095
34 (63.0%)
8 (32.0%)
Time to nadir PSA (ROC cut-off value)
≤6.5
36 (66.7%)
5 (20.0%)
>6.5
18 (33.3%)
20 (80.0%)
PSA doubling time (ROC cut-off value)
≤5.05
43 (79.6%)
4 (16.0%)
>5.05
11 (20.4%)
21 (84.0%)
PSA velocity (ROC cut-off value)
≤0.55
15 (27.8%)
23 (92.0%)
>0.55
39 (72.2%)
2 (8.0%)
PSA in the moment of disease reevaluation (ROC cut-off value)
